Skip to main content
Back to challenges
Challenge

Real-Time Localized E-commerce Assistant

Develop a cutting-edge real-time e-commerce assistant designed to facilitate the international expansion of grocery delivery services. This challenge focuses on generating dynamic, localized product descriptions, marketing copy, and providing a voice-enabled shopping experience for new markets like Spain and France. The assistant must intelligently adapt content based on regional linguistic nuances and cultural preferences, utilizing the Vercel AI SDK for its streaming capabilities and efficient interaction with large language models. Participants will build a serverless AI backend that interacts with a product catalog stored in AWS DynamoDB and leverages Claude 3.5 Haiku for rapid text generation and translation. The frontend interaction will demonstrate a voice-driven user interface using VAPI, showcasing how modern generative AI can create seamless, personalized e-commerce experiences across diverse global markets. Emphasis is placed on creating responsive, scalable, and culturally aware AI solutions.

Workflow AutomationHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Build a real-time localized e-commerce assistant using Vercel AI SDK, Claude 3.5 Haiku, AWS services, and VAPI for voice interaction.

Delivery guide

How work is evaluated

Evaluation

The evaluation will assess the assistant's ability to generate accurate, localized, and contextually appropriate e-commerce content, its real-time responsiveness, and the effectiveness of its voice interface.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
4 scoring checks
Binary
4 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

CorrectLanguageOutput

Ensures generated content is in the requested market's primary language.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

FeatureInclusion

Verifies that all specified product features are mentioned or implied in the description.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 3

Localization_Accuracy

BLEU score or similar for translation quality and cultural appropriateness. • target: 0.8 • range: 0.5-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 4

Voice_Intent_Accuracy

Percentage of correctly recognized voice intents. • target: 0.9 • range: 0.7-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Master the Vercel AI SDK for building streaming AI interfaces and connecting to various LLM providers in a serverless environment.

  • Implement real-time localization strategies for e-commerce content, including product descriptions and marketing copy, leveraging Claude 3.5 Haiku's speed and linguistic capabilities.

  • Design and integrate a voice user interface using VAPI to enable natural language shopping experiences, handling intent recognition and dynamic response generation.

  • Build a scalable data backend on AWS (e.g., DynamoDB for product catalog, S3 for media assets) to support multilingual content and user interactions.

  • Develop API routes and serverless functions using Next.js/Vercel platform to handle AI inference, tool calling, and data persistence for the e-commerce assistant.

  • Orchestrate intelligent content adaptation logic that considers regional customs, currency, and promotional strategies for different geographical markets.

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 2 tasks

Challenge setup

The evaluation will assess the assistant's ability to generate accurate, localized, and contextually appropriate e-commerce content, its real-time responsiveness, and the effectiveness of its voice interface.

Challenge input

{"product_name": "string", "product_features": ["string"], "market": "string"}

Agent execution

No agent tool is listed for this challenge.

Evaluated output

{"localized_description": "string", "language": "string"}

Evaluation checks
  • Ensures generated content is in the requested market's primary language.
  • Verifies that all specified product features are mentioned or implied in the description.
  • BLEU score or similar for translation quality and cultural appropriateness. • target: 0.8 • range: 0.5-1
Configured evaluation evidence
  • Localization Accuracy target: 0.8
  • 2 public reference cases
Configured runtime
  • Python execution harness

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

No tool records are attached.

Evaluation contract

  • CorrectLanguageOutput · Weight 1
  • FeatureInclusion · Weight 1
  • Localization_Accuracy · Weight 1
  • Voice_Intent_Accuracy · Weight 1

Metrics

  • Localization_Accuracy · Target 0.8
  • Voice_Intent_Accuracy · Target 0.9

Reference cases

  • 2 public reference cases

Workflow tasks

  1. Task 1

    Input: {"product_name": "string", "product_features": ["string"], "market": "string"}

    Output: {"localized_description": "string", "language": "string"}

    Execution harness: Python

  2. Task 2

    Input: {"audio_transcript": "string", "user_intent": "string", "context": "object"}

    Output: {"audio_response_transcript": "string", "detected_action": "string"}

    Execution harness: Python

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Real-Time Localized E-commerce Assistant